AWS Executive: Use Business Metrics to Measure AI

AWS Senior Vice President Colleen Aubrey argues companies should stop inventing new AI metrics and instead measure AI initiatives against existing business metrics like working capital and operational efficiency. Drawing on 20-plus years at Amazon, she outlines a practical playbook built around the Amazon Connect product family. Key lessons include avoiding feature parity traps by starting narrow, tying AI projects to CFO-tracked metrics, and recognizing that AI moves bottlenecks rather than eliminating them. Wells Vehicle Electronics is cited as an example, cutting inventory 7% while expanding monitored SKUs from 20 to 2,000 in three months.